The original target sample size was 16,000 households. Half of the sample (8,000) was generated by oversampling with five independent sampling frames (called “large provinces”) of Shanghai, Liaoning, Henan, Gansu, and Guangdong. Each of the subsample had 1,600 households. The other half 8,000 households were from an independent sampling frame composed of 20 provinces (called “small provinces”). The “large provinces” were representative at the regional level, which could contribute to provincial population inferences and cross-region comparisons. With a second-stage sampling, the five “large provinces”, together with “small provinces”, made up the overall sampling frame representative of the national population.

Taking the reginal differences in Chinese society and the reduction of survey processing cost into consideration, CFPS implemented Probability-Proportional-to-Size Sampling (PPS) with implicit stratification. Administrative units and socioeconomic status (SES) are used as the main stratification variables. Within the administrative unit, local GDP per capita was used as the ordering index for SES. If the GDP per capita in the administrative unit is not available, the proportion of nonagricultural population or population density is used alternatively.

All the sub-sampling frames of CFPS were obtained through three stages: the Primary Sampling Unit (PSU) was administrative districts/counties, the Second-stage 18 Sampling Unit (SSU) was administrative villages/neighborhood communities, and the third-stage (Ultimate) Sampling Unit (TSU) was households. In the first and Second stages, CFPS used official administrative divisions for the sample selection. The third sampling stage is a systematic selection of housing units from street listing with random starting point and equal probability method.

For detailed information of sampling design, Please visit CFPS Web site.